Application Development¶
Phase 4 — Track B — Nvidia Jetson · Module 5 of 7
Focus: Build production application software on the Jetson Orin Nano 8GB — from low-level peripheral access (GPIO, UART, SPI, I2C, CAN) through networking, GUI, multimedia pipelines, ML/AI inference (TensorRT, DLA, tinygrad), and ROS 2 integration.
Primary hardware: Jetson Orin Nano 8GB on custom or dev-kit carrier
Previous: 4. FSP Customization · Next: 6. Security and OTA
Sub-modules¶
| # | Sub-module | Focus |
|---|---|---|
| 1 | Peripheral Access | GPIO, PWM, UART, SPI, I2C, CAN, USB, NVMe/SD, backlight — Linux userspace and kernel driver interfaces |
| 2 | Network and Connectivity | Ethernet, Wi-Fi, Bluetooth, VPN, web server — wired and wireless connectivity stack |
| 3 | GUI | Qt, LVGL, web-based UI, framebuffer — display and touch for embedded products |
| 4 | Multimedia | Audio, cameras (USB + CSI), GStreamer, display output, video encode/decode — hardware-accelerated media pipelines |
| 5 | ML and AI | Quantization, TensorRT, DLA, tinygrad, DeepStream, profiling — edge inference optimization |
| 6 | ROS 2 | Nodes, Nav2, DDS, real-time, Jetson deployment — robot software framework |
How the sub-modules fit the product flow¶
Custom carrier board (Module 2) + L4T BSP (Module 3) + FSP (Module 4)
│
├─ Peripheral Access ─── talk to sensors, actuators, buses
├─ Network ──────────── connect to cloud, fleet, local network
├─ GUI ──────────────── user-facing display (if applicable)
├─ Multimedia ────────── cameras, audio, video pipelines
├─ ML / AI ──────────── inference optimization on Orin Nano
└─ ROS 2 ────────────── robot middleware tying it all together
│
▼
Security & OTA (Module 6) → Compliance (Module 7)
Work through sub-modules 1–4 in any order based on your product's needs. ML/AI (sub-module 5) and ROS 2 (sub-module 6) typically build on the peripheral and multimedia foundations.